HFLM Python integration way slower than CLI

Models
2

This may help avoid some of the problems.

eval_model = HFLM(
    pretrained="models/Llama-2-7b-hf",
    dtype="bfloat16",
    device="cuda",
    #device_map="auto",
)
results = evaluator.simple_evaluate(
    model=eval_model,
    tasks=["arc_easy","arc_challenge","hellaswag","lambada_openai","lambada_standard","openbookqa","piqa","winogrande"],
    batch_size=64,
    max_batch_size=64,
)